Neuron Selection for RBF Neural Network Classifier Based on Data Structure Preserving Criterion

Effective feature selection is essential to the development of any intelligent classifier which is intended for use in high-dimension domains. This paper presents an approach that incorporates a rough set-assisted feature reduction method and a neural network-based classifier for image classification. The approach minimises the need for feature extraction without altering the underlying semantics of the features chosen. Through the proposed integration the size of the neural network classifier, which is sensitive to the dimensionality of the dataset, becomes manageable and the network is able to classify images that would otherwise require many more features to represent. Comparative study results from realistic applications demonstrate the success of this work.

Download Full-text

A POD-Based Center Selection for RBF Neural Network in Time Series Prediction Problems

Adaptive and Natural Computing Algorithms - Lecture Notes in Computer Science ◽

10.1007/978-3-540-71629-7_22 ◽

2007 ◽

pp. 189-198 ◽

Cited By ~ 5

Author(s):

Wenbo Zhang ◽

Xinchen Guo ◽

Chaoyong Wang ◽

Chunguo Wu

Keyword(s):

Neural Network ◽

Time Series ◽

Rbf Neural Network ◽

Time Series Prediction ◽

Selection For ◽

Prediction Problems

Download Full-text

A Self-Adaptive RBF Neural Network Classifier for Transformer Fault Analysis

IEEE Transactions on Power Systems ◽

10.1109/tpwrs.2010.2040491 ◽

2010 ◽

Vol 25 (3) ◽

pp. 1350-1360 ◽

Cited By ~ 71

Author(s):

Ke Meng ◽

Zhao Yang Dong ◽

Dian Hui Wang ◽

Kit Po Wong

Keyword(s):

Neural Network ◽

Rbf Neural Network ◽

Fault Analysis ◽

Neural Network Classifier ◽

Transformer Fault ◽

Self Adaptive

Download Full-text

A Variable Selection for Asphalt Pavement Performance Based on RBF Neural Network

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.744-746.1222 ◽

2015 ◽

Vol 744-746 ◽

pp. 1222-1225

Author(s):

Peng Tian ◽

Gao Feng Zhan ◽

Lei Nai

Keyword(s):

Neural Network ◽

Variable Selection ◽

Rbf Neural Network ◽

Asphalt Pavement ◽

Asphalt Mixture ◽

Pavement Performance ◽

Method Selection ◽

Selection Of Variables ◽

Selection For ◽

Selection Of

By combining RBF neural network with MIV algorithm, the main influencing factors of asphalt mixture pavement performance will be selected. First, the MIV values will be calculated by MIV method. Selection of variables is based on the size of MIV. There are 8 variables selected form 12 variables. Then, a new RBF neural network will be found by the data which have great impact to the output result. The comparison between the two RBF simulate results will prove that the method of MIV is feasible in variable selection. By the MIV method, the simulate results of RBF will be calculated faster and more accurately.

Download Full-text